Objective: Although gestational diabetes is among the most common diseases arising during pregnancy, glucose stix is the only screening test to date in Germany. Our goal was to evaluate the sensitivity of the glucose-stix for diabetes screening and the possible influence of other parameters.
Methods: 1001 patients who underwent the 50 g glucose screening test between June 1, 1997 and January 5, 2000 as part of prenatal care were asked to participate. In accordance with the guidelines of the American Diabetes Association, patients with a screening test result >/= 140 mg/dl underwent a oral glucose tolerance test (Carpenter/Coustan criteria). A urine sample was collected prior to the test. The glucose content of the urine was semiquantitatively analyzed using a test strip (Multistix 10 SG Bayer), Munich, Germany). Blood pressure was measured in 349 consecutive cases according to the criteria of the National Institute of Health.
Results: The overall frequency of gestational diabetes was 4.1% (37/912). 8.2% of the women presented with glycosuria (82/1001, 36 before screening, 46 based on the pregnancy medical records booklet). 30/82 (37%) of these patients had a pathological screening test (P = 0.029). 7.1% (52/729) of the healthy patients and 10.8% (4/37) of the gestational diabetics had glycosuria at least once. Therefore, the sensitivity of glycosuria is 10.8%, the positive predictive value is 6.6%. The systolic blood pressure was 116+/- 12 mmHg and the diastolic blood pressure 72 +/- 9 mmHg. Three of 349 (0.9%) patients were documented with preexisting hypertension, 14/349 (4.0%) patients with "pregnancy induced hypertension". Patients with glycosuria were both significantly more advanced in gestational age (34.4 +/- 2.8 versus 33.7 +/- 2.9, P = 0.673) and had higher diastolic blood pressure (79 +/- 9 versus 71 +/- 9, P = 0.005). The 50 g glucose screening test results showed only a tendency to differ (131 +/- 23 versus 127 +/- 24, P = 0.073). A multivariate analysis of these factors showed a significant influence of the diastolic blood pressure (P = 0.016) and the 50 g glucose screening test (P = 0.032), whereas the gestational week had no influence (P = 0.673).
Conclusions: Urine glucose dip stick analysis is not useful in the detection of gestational diabetes because of its low sensitivity and negative predictive value. Our study suggests that glycosuria is not only dependent on the blood glucose level, but highly influenced by diastolic blood pressure. The results clearly underscore the need for standardized, routine testing of every pregnant woman.